fix local pydoc generation

fix relative pydoc links

docs/website/blog/2024-01-15-dlt-dbt-runner-on-cloud-functions.md

@@ -150,7 +150,7 @@ The second option is running dbt using data load tool(dlt).
 
 I work at dlthub and often create dlt pipelines. These often need dbt for modeling the data, making
 the dlt-dbt combination highly effective. For using this combination on cloud functions, we used
-[dlt-dbt runner](https://dlthub.com/docs/api_reference/helpers/dbt/runner#create_runner) developed
+[dlt-dbt runner](../../api_reference/helpers/dbt/runner#create_runner) developed
 at dlthub.
 
 The main reason I use this runner is because I load data with dlt and can re-use dlt’s connection to
@@ -331,7 +331,7 @@ event-driven pipelines with small to medium workloads. For larger data loads nea
 consider separating dlt and dbt into different cloud functions.
 
 > For more info on using `dlt-dbt runner` , please refer to the
-> [official documentation by clicking here.](https://dlthub.com/docs/api_reference/helpers/dbt/runner#dbtpackagerunner-objects)
+> [official documentation by clicking here.](../../api_reference/helpers/dbt/runner#dbtpackagerunner-objects)
 
 ### Deployment considerations: How does cloud functions compare to Git Actions?
 

docs/website/docs/dlt-ecosystem/destinations/bigquery.md

@@ -263,7 +263,7 @@ bigquery_adapter(my_resource, partition="partition_column_name")
 my_resource = bigquery_adapter(my_resource, partition="partition_column_name")
 ```
 
-Refer to the [full API specification](https://dlthub.com/docs/api_reference/destinations/impl/bigquery/bigquery_adapter) for more details.
+Refer to the [full API specification](../../api_reference/destinations/impl/bigquery/bigquery_adapter) for more details.
 
 <!--@@@DLT_TUBA bigquery-->
 

docs/website/package.json

@@ -4,7 +4,7 @@
   "private": true,
   "scripts": {
     "docusaurus": "docusaurus",
-    "start": "node tools/update_version_env.js && node tools/preprocess_docs.js && concurrently --kill-others \"node tools/preprocess_docs.js --watch\" \"docusaurus start\"",
+    "start": "PYTHONPATH=. poetry run pydoc-markdown && node tools/update_version_env.js && node tools/preprocess_docs.js && concurrently --kill-others \"node tools/preprocess_docs.js --watch\" \"docusaurus start\"",
     "build": "node tools/preprocess_docs.js  && PYTHONPATH=. poetry run pydoc-markdown && node tools/update_version_env.js && docusaurus build",
     "build:netlify": "node tools/preprocess_docs.js && PYTHONPATH=. pydoc-markdown && node tools/update_version_env.js && docusaurus build --out-dir build/docs",
     "swizzle": "docusaurus swizzle",